protected virtual void LoadRange(int aRange)
{
nextElement = set.GetRangeStart(aRange);
endElement = set.GetRangeEnd(aRange);
}
//------------------------------------------------------------------------
//
// build Build the list of non-overlapping character ranges
// from the Unicode Sets.
//
//------------------------------------------------------------------------
internal virtual void Build()
{
RangeDescriptor rlRange;
if (fRB.fDebugEnv != null && fRB.fDebugEnv.IndexOf("usets", StringComparison.Ordinal) >= 0)
{
PrintSets();
}
// Initialize the process by creating a single range encompassing all characters
// that is in no sets.
//
fRangeList = new RangeDescriptor();
fRangeList.fStartChar = 0;
fRangeList.fEndChar = 0x10ffff;
//
// Find the set of non-overlapping ranges of characters
//
foreach (RBBINode usetNode in fRB.fUSetNodes)
{
UnicodeSet inputSet = usetNode.fInputSet;
int inputSetRangeCount = inputSet.RangeCount;
int inputSetRangeIndex = 0;
rlRange = fRangeList;
for (; ;)
{
if (inputSetRangeIndex >= inputSetRangeCount)
{
break;
}
int inputSetRangeBegin = inputSet.GetRangeStart(inputSetRangeIndex);
int inputSetRangeEnd = inputSet.GetRangeEnd(inputSetRangeIndex);
// skip over ranges from the range list that are completely
// below the current range from the input unicode set.
while (rlRange.fEndChar < inputSetRangeBegin)
{
rlRange = rlRange.fNext;
}
// If the start of the range from the range list is before with
// the start of the range from the unicode set, split the range list range
// in two, with one part being before (wholly outside of) the unicode set
// and the other containing the rest.
// Then continue the loop; the post-split current range will then be skipped
// over
if (rlRange.fStartChar < inputSetRangeBegin)
{
rlRange.Split(inputSetRangeBegin);
continue;
}
// Same thing at the end of the ranges...
// If the end of the range from the range list doesn't coincide with
// the end of the range from the unicode set, split the range list
// range in two. The first part of the split range will be
// wholly inside the Unicode set.
if (rlRange.fEndChar > inputSetRangeEnd)
{
rlRange.Split(inputSetRangeEnd + 1);
}
// The current rlRange is now entirely within the UnicodeSet range.
// Add this unicode set to the list of sets for this rlRange
if (rlRange.fIncludesSets.IndexOf(usetNode) == -1)
{
rlRange.fIncludesSets.Add(usetNode);
}
// Advance over ranges that we are finished with.
if (inputSetRangeEnd == rlRange.fEndChar)
{
inputSetRangeIndex++;
}
rlRange = rlRange.fNext;
}
}
if (fRB.fDebugEnv != null && fRB.fDebugEnv.IndexOf("range", StringComparison.Ordinal) >= 0)
{
PrintRanges();
}
//
// Group the above ranges, with each group consisting of one or more
// ranges that are in exactly the same set of original UnicodeSets.
// The groups are numbered, and these group numbers are the set of
// input symbols recognized by the run-time state machine.
//
// Numbering: # 0 (state table column 0) is unused.
// # 1 is reserved - table column 1 is for end-of-input
// # 2 is reserved - table column 2 is for beginning-in-input
// # 3 is the first range list.
//
RangeDescriptor rlSearchRange;
for (rlRange = fRangeList; rlRange != null; rlRange = rlRange.fNext)
{
for (rlSearchRange = fRangeList; rlSearchRange != rlRange; rlSearchRange = rlSearchRange.fNext)
{
if (ListEqualityComparer <RBBINode> .Default.Equals(rlRange.fIncludesSets, rlSearchRange.fIncludesSets))
{
rlRange.fNum = rlSearchRange.fNum;
break;
}
}
if (rlRange.fNum == 0)
{
fGroupCount++;
rlRange.fNum = fGroupCount + 2;
rlRange.SetDictionaryFlag();
AddValToSets(rlRange.fIncludesSets, fGroupCount + 2);
}
}
// Handle input sets that contain the special string {eof}.
// Column 1 of the state table is reserved for EOF on input.
// Column 2 is reserved for before-the-start-input.
// (This column can be optimized away later if there are no rule
// references to {bof}.)
// Add this column value (1 or 2) to the equivalent expression
// subtree for each UnicodeSet that contains the string {eof}
// Because {bof} and {eof} are not a characters in the normal sense,
// they doesn't affect the computation of ranges or TRIE.
string eofString = "eof";
string bofString = "bof";
foreach (RBBINode usetNode in fRB.fUSetNodes)
{
UnicodeSet inputSet = usetNode.fInputSet;
if (inputSet.Contains(eofString))
{
AddValToSet(usetNode, 1);
}
if (inputSet.Contains(bofString))
{
AddValToSet(usetNode, 2);
fSawBOF = true;
}
}
if (fRB.fDebugEnv != null && fRB.fDebugEnv.IndexOf("rgroup", StringComparison.Ordinal) >= 0)
{
PrintRangeGroups();
}
if (fRB.fDebugEnv != null && fRB.fDebugEnv.IndexOf("esets", StringComparison.Ordinal) >= 0)
{
PrintSets();
}
fTrie = new Trie2Writable(0, // Initial value for all code points.
0); // Error value for out-of-range input.
for (rlRange = fRangeList; rlRange != null; rlRange = rlRange.fNext)
{
fTrie.SetRange(
rlRange.fStartChar, // Range start
rlRange.fEndChar, // Range end (inclusive)
rlRange.fNum, // value for range
true // Overwrite previously written values
);
}
}
internal virtual void LoadRange(int aRange) // ICU4N specific - marked internal instead of protected, since the functionality is obsolete
{
nextElement = set.GetRangeStart(aRange);
endElement = set.GetRangeEnd(aRange);
}